CREATION BY EVOLUTION
very slowly and imperceptibly. Probably the same was true of the ants of periods antedating the Tertiary, though there may have been in those periods occasional spells of acceleration and efflorescence of new forms.
When we carefully study the anatomy and development of the various species of ants we find that they are essentially wasps, and that they are closely allied to species of certain existing families of wasps, the Tiphiidae, Mutillidae, and Thynnidae. We must, indeed, suppose that the ancestors of these families produced also the ants, the Formicidae. But the members of these families, like most other wasps, are solitary, and, like most animals, possess only a single type of female; whereas among the ants each species presents two female phases, or castes, one of which, the “queen,” is fertile and nearly always winged, and the other, the “worker,” is always wingless and nearly always sterile. In only a few species of ants, and those highly parasitic species, do we find no worker caste. There is every reason to assume that in these species the worker has been lost or suppressed within comparatively recent time. We must therefore conclude that sexual trimorphism—that is, the presence in each species of three castes, male, fertile female, and sterile female, or worker, which were perfectly developed also in the known fossil ants of Tertiary time—was first established among the Mesozoic ancestors of the family Formicidae. A similar trimorphism has arisen independently among the social bees and social wasps, but it has evidently been of much more recent development, for among these insects the worker is much more like the fertile female and always has wings. Then, too, the differentiation of fertile and sterile females among certain tropical wasps is so feeble that the evolution of the two castes may be said to be still uncompleted.
When we arrange all the species of living and fossil ants
[ 216 ]
